Amazon Redshift

AWS cloud data warehouse with SQL-based machine learning built in

AI Data AnalysisOverseasβ˜…β˜…β˜…β˜†β˜† 3.0

What is Amazon Redshift?

Amazon Redshift is AWS's petabyte-scale data warehouse. It supports columnar storage, massively parallel query execution, workload management for mixed concurrency and Spectrum queries that reach data in S3 without loading it first. Redshift ML lets analysts create models using SQL, with SageMaker handling training behind the scenes, and provides prediction functions callable from queries. Both provisioned clusters and a serverless option are available, so teams can choose between reserved capacity and pay-per-use. It is a mature warehouse first and an AI feature set second, which is the right way to evaluate it.

Last updated: 2026-09-20. This site only provides an index; for exact features, pricing, and licensing, see the official website.

Key features

  • Columnar storage with massively parallel query execution
  • Redshift ML for training and predicting with SQL and SageMaker
  • Spectrum queries that read S3 data without loading it
  • Workload management to separate ETL, BI and ad hoc traffic
  • Provisioned clusters or a serverless pay-per-use option
  • Materialised views, concurrency scaling and automated snapshots

Pros & cons

Strengths

  • Very large datasets stay queryable with predictable performance
  • Redshift ML keeps simple modelling inside familiar SQL
  • Serverless option removes cluster sizing guesswork

Watch out for

  • Costs grow with nodes or usage and need active monitoring
  • Tuning distribution and sort keys requires real expertise
  • Tight AWS coupling makes multi-cloud use awkward

Best for & use cases

large-scale warehousing, bi backends, etl consolidation and sql-based prediction

If you're comparing similar products, check the alternatives below, or browse all tools in the AI Data Analysis category.

FAQ

Redshift or Athena?

Redshift suits repeated, complex, high-concurrency workloads where consistent performance justifies provisioned capacity. Athena suits occasional queries over files already in S3 with no infrastructure. Many organisations use both and keep the expensive queries on Redshift.

Is Redshift ML a real machine learning platform?

It is a convenience layer: you write SQL to create a model and SageMaker trains it, then you call a prediction function. It covers common tabular problems well, but serious model development still happens in a dedicated ML environment.

What drives the bill?

Node hours or serverless compute, storage, and the data scanned by Spectrum queries. Idle clusters and unpartitioned S3 scans are the two most common sources of avoidable cost.